Search

ObjectiveTo estimate dose-response associations between non-occupational physical activity and multiple chronic disease outcomes in the general adult population.Eligibility criteriaProspective cohort studies with (a) general population samples >10,000 adults, (b) ≥3 exposure categories, and (c) risk measures and confidence intervals for all-cause mortality, total cardiovascular disease, coronary heart disease, stroke, heart failure, total cancer, and site-specific cancers (head and neck, myeloid leukemia, myeloma, gastric cardia, lung, liver, endometrium, colon, breast, bladder, rectum, esophagus, prostate, kidney).Information sourcesPubMed, Scopus, Web of Science, and reference lists of published studies, searched in February 2019.Data extraction and synthesisIndependent extraction and double-checking of study characteristics, exposure, and outcome assessment by two reviewers for each paper. Primary exposure was non-occupational physical activity volume, harmonized to physical activity energy expenditure in marginal MET-hours per week (mMET-h/week). The current minimum recommendations for physical activity (150 min/week of moderate-to-vigorous physical activity) equate to 8.75 mMET-h/week. Outcomes were risks of mortality, cardiovascular diseases, and cancers. We used restricted cubic splines in random-effects meta-analyses. Potential population impact was quantified using impact fractions.Results196 articles were included, covering 94 cohorts. The evidence base was largest for all-cause mortality (50 independent results; 163,415,543 person-years; 811,616 events), and incidence of cardiovascular disease (37 independent results; 28,884,209 person-years; 74,757 events) and cancer (31 independent results; 35,500,867 person-years; 185,870 events). In general, inverse non-linear associations were observed, steeper between 0 and 8.75 mMET-h/week, with smaller marginal reductions in risk above this level to 17.5 mMET-h/week, beyond which additional reductions were small and uncertain. Associations were stronger for all-cause and cardiovascular disease mortality than for cancer mortality. If all insufficiently active individuals had met the recommended physical activity level, 15.7% (95%CI: 13.1 to 18.2%) of all premature deaths would have been averted.ConclusionsInverse non-linear dose-response associations suggest substantial protection against a range of chronic disease outcomes from small increases in non-occupational physical activity in inactive adults.Review registrationPROSPERO CRD42018095481.

ObjectivesTo investigate associations between occupational physical activity (OPA) and all-cause mortality.MethodsFrom baseline (2006-2010), 452,884 UK Biobank participants (aged 40-69 years) were followed for a median 11.1 (IQR: 10.4-11.8) years. OPA was categorised by cross-tabulating degree of manual work and walking/standing work amongst those in paid employment (n=264,424), whereas categories of occupational status were used for those not in paid employment (n=188,460). Cox proportional hazards models were used to estimate hazard ratios (HR) and 95% confidence intervals (CI) for all-cause mortality by occupational category, and for working hours/week and non-occupational physical activity stratified by occupational category.ResultsDuring 4,965,616 person-years of follow-up, 23,310 deaths occurred. Compared to those in sedentary jobs, retirement was associated with lower mortality in women (HR=0.74, CI:0.68-0.81) and men (HR=0.85, CI:0.79-0.92), whereas unemployment was associated with higher mortality in men (HR=1.26, CI:1.10-1.45). There was no evidence of mortality differences by OPA category within the working population. Working ConclusionWork classified as having higher levels of OPA may not be as active as reported, or the types of physical activity performed in those jobs are not health-enhancing. Irrespective of OPA category or employment status, non-occupational physical activity appears to provide health benefits.SUMMARY BOXWhat are the new findings?Retirement was associated with lower all-cause mortality, compared to having a sedentary job in both men and women but unemployment was only associated with higher mortality in men.There were no differences in mortality between categories with different levels of self-reported OPAPhysical activity outside of work was associated with lower hazard of all-cause mortality and there was no interaction with occupational physical activity, indicating similar benefits across different jobs types.How might it impact on clinical practice in the future?Health professionals should be aware that occupations assumed to involve more physical activity may not be as active as reported, or the types of physical activity performed in those jobs may not be health-enhancing.Physical activity during leisure-time should continue to be recommended to adults of all paid and unpaid occupational categories.

Harmonisation of data for pooled analysis relies on the principle of inferential equivalence between variables from different sources. Ideally, this is achieved using models of the direct relationship with gold standard criterion measures, but the necessary validation data are often unavailable. This study examines an alternative method of harmonisation by indirect validation. Starting methods were self-report or accelerometry, from which we derived indirect models of relationships with doubly labelled water (DLW)-measured physical activity energy expenditure (PAEE) using sets of two bridge equations via one of three intermediate measures. Coefficients and performance of indirect models were compared to corresponding direct validation models (linear regression of DLW-measured PAEE on starting methods). Indirect model beta coefficients were attenuated compared to direct model betas (10-63%), narrowing the range of PAEE values; attenuation was greater when bridge equations were weak. Directly and indirectly harmonised models had similar error variance but most indirectly derived values were biased at group-level. Correlations with DLW-measured PAEE were identical after harmonisation using continuous linear but not categorical models. Wrist acceleration harmonised to DLW-measured PAEE via combined accelerometry and heart rate sensing had lowest error variance (24.5%) and non-significant mean bias 0.9 (95%CI: −1.6; 3.4) kJ•day−1•kg−1. Associations between PAEE and BMI were similar for directly and indirectly harmonised values, but most fell outside the confidence interval of the criterion PAEE-to-BMI association. Indirect models can be used for harmonisation. Performance depends on the measurement properties of original data, variance explained by available bridge equations, and similarity of population characteristics.